At the Chicago Tribune where I worked for 32 years, the story of Bill Recktenwald and the parking meters was legendary — and instructive.
From 1984 to early 1987, Chicago suffered a huge loss in parking meter revenue — 42 percent — and city officials couldn’t explain it. Complaints about broken meters had nearly doubled in a three-year span, but officials said only 20 percent of the city’s 29,278 meters were broken.
So, Recktenwald, my hard-nosed colleague, set out to create his own database. Over a two-week period, he walked with a lot of small change from block to block around the city and, as he reported to Tribune readers on January 25, 1987, checked 1,039 meters (3.5 percent of the total).
Some, he found clearly broken. Some wouldn’t accept a coin. Some took coins but gave no time or less time on the meter than was purchased — and some gave more time.
In all, he found that 70 percent of the meters didn’t work properly, so, if his findings were at all representative of what was going on across the city, it was no wonder Chicago was getting less revenue than it should have.
This was a lesson to the rest of us at the Tribune on the value of creating a database and on the rich, deep insight that such can provide in telling what the facts are.
New worlds of questions
Recktenwald researched and wrote this story at a time when journalists were just starting to understand the revelations that such reporting could discover. Previously, reporters had relied on data collected and analyzed by government officials or by outsiders, such as academics and public interest groups, all of which was and remains useful.
The idea of creating your own database — Recktenwald’s was a small, shoe-leather version of this approach — opened up new worlds of questions that could be raised and answered. Not all reporters bought into this, but, for those who did, the result was a wealth of unexpected and often impactful stories.
I did my own versions of this by taking census data for Chicago or the metropolitan region or the state and matching the numbers up against other statistics, such as crime reports, housing information and even telephone usage. Eventually, I was able to tap into the expertise of computer-savvy colleagues to help me develop more complex research.
Today, of course, major news organizations, such as the New York Times and the Washington Post, routinely crunch numbers to find the story behind the story.
Data-driven history of art
I mention all of this because Diana Seave Greenwald, assistant curator at the Isabella Stewart Gardner Museum in Boston, is playing a Bill Recktenwald-like role in the art world with her newly published Painting by Numbers: Data-Driven Histories of Nineteenth-Century Art (Princeton University Press, 256 pages, $35).
Let me be quick to make clear that Greenwald’s book, adapted, in part, from her 2018 Ph.D. dissertation at the University of Oxford, is much more sophisticated than Recktenwald’s stroll around Chicago feeding quarters into parking meters.
Indeed, Painting by Numbers is an important and revelatory work that opens a door to new and unprecedented ways of doing art history research.
Greenwald, an expert in art history and economic history, has developed a shiny, nifty and highly useful tool for a blending of both disciplines, an approach that she calls the data-driven history of art. It is, she writes, “a way to unite art and economic history” while building on the digital techniques that are already in use in the humanities.
Her book “aims to show that one can engage with both the empirical quantitative methods of modern economics and its theories about the structural complexities of human behavior and institutions.”
In other words, that the computational tools of economic history, such as regression analysis, as well as the theories of economics, can complement and enhance the work of art historians.
“Berated me loudly”
The idea of a data-driven history of art is not an easy sell.
In fact, on the first page of Painting by Numbers, Greenwald tells the story of art historian Jules Prown being jeered in the 1960s when he presented a paper on his use of the computer in analyzing the paintings and patronage of American Painter John Singleton Copley. “After the conclusion of the session, a senior art historian berated me loudly in the aisle for my apostasy,” Prown reported.
Art history, Greenwald notes, has long been based on a focused study of select artists and art works and on approaching those works with close looking, that is, the effort to see all that is contained in, say, a painting, from its brushstrokes to its colors to its images. By contrast, economics and other social sciences seek a “zoomed-out view of history” through the use and analysis of large quantitative datasets.
This is an approach, she recognizes, that isn’t appropriate in all cases:
The hypothetical scholar conducting a close study of the work of Edgar Degas…probably does not need large datasets, statistics, or economic theory to guide him or here.
However, when the scholar becomes interested in how Degas and his work fit into broader trends in the art world or — as in the case of the social history of art — broader trends in society, the data-driven history of art becomes useful.
And it isn’t an approach to replace the long-used techniques of art history:
A close view of particular artworks and artists will always be necessary for art historical scholarship. However, it is valuable to also consider the macroscopic view of artistic trends, as revealed by data-driven methods of research.
The use of all these approaches, Greenwald argues, will enable scholars to examine art “in more complete socioeconomic and cultural contexts.”
Three case studies
To illustrate the data-driven history of art in action, Greenwald presents three case studies of the 19th century art world:
- Rural genre paintings in France: Scholars have suggested that an apparent rise in the number of artworks of peasant scenes in the countryside in the 1800s was due to the increasing industrialization of the cities. But Greenwald’s crunching of the numbers, instead, seems to point to two other factors as more important: cheaper rail travel to rural areas and the establishment of artist colonies.
- The low proportion of artwork by women in museums: In addition to the clear cultural and social strictures that held women back, Greenwald’s examination of the data indicates that women, pressed for time as householders and mothers, tended to work on art that could be done in fits and starts, such as still lives and works on paper, unlike the sort with much higher esteem among buyers and collectors: landscapes, portraits and history works that required long time commitments. This became a double-whammy later for women artists, according to Greenwald. “Not only are [the works of women] less present in museums, but they are most present in light-sensitive media that can spend limited time on display.”
- The Empire as the subject matter of art at the annual Royal Academy exhibitions: The British of the 19th and early 20th century took great pride in their empire, and, yet, when Greenwald examined the data of 145 years of Royal Academy exhibitions, she found to her amazement that the nations of the empire were rarely depicted. Her conclusion is that the art of the exhibitions was essentially a celebration of the greater freedoms and stability of the British society during those years. The colonies were ignored, she suggests, because the exact opposite was happening there — slavery, instability and victimization.
“A bird’s-eye view”
Greenwald’s Painting by Numbers is essentially a how-to book.
She is at pains to explain how she found and, in one case, created the main databases she uses, and to explain the methods and theories of economic history that she employs, and to show chart after chart and table after table to illustrate her findings — and all the while pointing out how useful this approach is.
In her conclusion, she writes:
With these datasets and some guidance from the social sciences about how to productively analyze them, it is possible to glimpse the structural constraints that formed the decisions and behaviors of artists and other actors. A bird’s-eye view of the historical art world emerges.
Ideally, art historians using Greenwald’s new tool won’t need to explain in such detail in their texts how they developed the insights that the tool brings them. That nuts-and-bolts information is better in an appendix or an end note.
“What to ask”
This new tool, though, like any tool, is only as good as the person wielding it.
Halfway through Painting by Numbers, Greenwald quotes British art historian T. J. Clark who, in 1974, wrote:
We need facts — about patronage, about art dealing, about the status of the artist, the structure of artistic production — but we need to know what to ask of the material.
That is the key. The techniques, old and new, are important, but the essential element of art history — and, indeed, all intellectual endeavor — is the questions that are asked.
We can’t get the answers we need unless we know what questions to ask.
Greenwald’s new tool gives art historians the ability to ask even more and better questions. No scholar could ask for more.
Patrick T. Reardon